This paper investigates the dynamic scheduling of an automated manufacturing workcenter at which jobs are processed in batches, and there is a constant changeover time between batches of different part types. The primary measures of schedule performance are mean flow time and mean tardiness. The dynamic scheduling problem is treated as a series of static problems which are solved on a rolling-horizon basis. Characteristics of the optimal solutions to the mean flow time and mean tardiness problems are developed, and an implicit enumeration approach to the mean tardiness problem is proposed. These results are used for constructing efficient scheduling procedures for the dynamic problem. We also derive the steady state relationship between workcenter utilization level, batch size and mean flow time for one and two part types. A simulation study extends this relationship to a larger number of part types.